Desmoplakin and periplakin genetically and functionally contribute to eosinophilic esophagitis

Abstract

Eosinophilic esophagitis (EoE) is a chronic allergic inflammatory disease with a complex underlying genetic etiology. Herein, we conduct whole-exome sequencing of a multigeneration EoE pedigree (discovery set) and 61 additional multiplex families with EoE (replication set). A series of rare, heterozygous, missense variants are identified in the genes encoding the desmosome-associated proteins DSP and PPL in 21% of the multiplex families. Esophageal biopsies from patients with these variants retain dilated intercellular spaces and decrease DSP and PPL expression even during disease remission. These variants affect barrier integrity, cell motility and RhoGTPase activity in esophageal epithelial cells and have increased susceptibility to calpain-14-mediated degradation. An acquired loss of esophageal DSP and PPL is present in non-familial EoE. Taken together, herein, we uncover a pathogenic role for desmosomal dysfunction in EoE, providing a deeper mechanistic understanding of tissue-specific allergic responses.

Fig. 1. Identification of DSP and PPL variants by whole-exome sequencing.

a Schema depicting the workflow for whole-exome sequencing (WES) filtering of rare variants in five patients with eosinophilic esophagitis (EoE) in the discovery set and confirmation in the replication set, as detailed in Methods. b Simplified pedigree of F430 (Details are in Supplementary Fig. 1). The arrowhead in the lower left corner indicates the proband and the slash indicates a deceased subject. “Not assessed” indicates the subject having GI symptoms but had never undergone an esophagogastroduodenoscopy. c Pedigrees of families with DSP or PPL variants in the replication set. Solid symbols indicate subjects with EoE, and open symbols indicate unaffected subjects. Arrowheads in the lower left corner indicate probands. For variant genotyping, “m” indicates the mutant DSP or PPL allele and “+” the reference allele. d Protein domain architectures and the location of amino acids predicted by mutations for DSP and PPL. The red diamond indicates a “hotspot” for mutations associated with cardiocutaneous disorders. EoE eosinophilic esophagitis, DSP desmoplakin, PPL periplakin, AA amino acid, MAF minor allele frequency, GI gastrointestinal, NA not assessed, DSC1 desmocollin 1, JUP junction plakoglobin, PKP1 plakophilin 1, IFs intermediate filaments, EVPL envoplakin, PRD plakin repeat domain, LD linker domain.

Fig. 2. Expression of DSP and PPL in patients with EoE.

a DSP (left) and PPL (middle) mRNA expression in esophageal biopsies from controls and patients with non-familial EoE (inactive and active). Each point represents an individual subject [(Control, n = 48; Inactive EoE, n = 51; Active EoE, n = 147); red data points represent patients with DSP or PPL variants (not included in the statistics)]. Statistics: DSP (Control vs. Inactive EoE, P > 0.9999; Control vs. Active EoE, P = 0.0007; Inactive EoE vs. Active EoE, P = 0.009); PPL (Control vs. Inactive EoE, P > 0.9999; Control vs. Active EoE, P < 0.0001; Inactive EoE vs. Active EoE, P < 0.0001). Correlation plot of DSP and PPL mRNA expression is also shown (right) (n = 246). Statistics: P < 0.0001. b Peak esophageal eosinophil counts (left) and TSLP mRNA expression (right) are plotted by groups for non-familial and familial EoE with DSP or PPL variants (non-familial EoE, n = 115; familial EoE with variants, n = 15). All samples were from the biopsies during the active disease state. The dashed line indicates the diagnostic threshold of EoE (15 eosinophil/hpf). Statistics: peak esophageal eosinophil count, P = 0.0927; TSLP, P = 0.0010. c Correlation plots of gene expressions (left: DSP, right: PPL) and histologic scores (red: eosinophil features, blue: structural features) (n = 68). Statistics: left (DSP with eosinophil features, P = 0.0022; DSP with structural features, P < 0.0001), right (PPL with eosinophil features, P = 0.0014; PPL with structural features, P < 0.0001). d and e Representative western blot analysis of DSP (d) and PPL (e) among control individuals (n = 4), patients with non-familial EoE (n = 4), and patients with familial EoE with DSP or PPL variants (n = 6). GAPDH serves as a loading control. For patients with EoE, all samples were from the biopsies during the inactive disease state. Statistics: d (control vs. non-familial EoE, P = 0.7833; control vs. familial EoE, P = 0.0022; non-familial EoE vs. familial EoE, P = 0.0007), e (control vs. non-familial EoE, P = 0.9061; control vs. familial EoE, P = 0.0317; non-familial EoE vs. familial EoE, P = 0.0142). For panels a (left and middle), b, d and e, data are presented as mean ± SEM. For panels a–e, n is the number of biologically independent subjects. For panels a–e, two-tailed P-values were determined by the following tests: a (left and middle), Kruskal–Wallis test followed by a Dunn multiple-comparison test; b, the unpaired t-test; a (right) and c, Spearman’s rank correlation coefficient (multiple comparisons were not applied); and d, e, one-way ANOVA test followed by a Tukey’s multiple comparisons test. *P < 0.05, **P < 0.01 and ***P < 0.001. EoE eosinophilic esophagitis, DSP desmoplakin, PPL periplakin, GAPDH Glyceraldehyde 3-phosphate dehydrogenase, hpf high-power microscopic field, IQR interquartile range, DAPI 4',6-diamidino-2-phenylindole, MW molecular weight, NS not significant, SEM standard error of the mean.

Fig. 3. Effects of EoE-associated DSP and PPL variants.

a Representative hematoxylin and eosin (H&E)–stained sections of EPC2 cells stably transduced with constructs encoding non-variant and mutated DSP or PPL after air–liquid interface (ALI) differentiation (day 14). Arrows point to the non-cellular areas that were formed. Scale bar: 50 μM. Data are representative of three experiments performed in duplicate. b The transepithelial electrical resistance (TEER) and c FITC-dextran flux measurements are shown for EPC2 cells grown at the ALI. d Wound healing assays performed in EPC2 cells transduced with constructs encoding non-variant and mutated DSP or PPL. Quantification of the wound closure after 8 h was shown. For panels b–d data are representative of three experiments performed in duplicate and are presented as mean ± SEM. Statistics (versus non-variants): b (p.G46D, P = 0.0324; p.R808C, P = 0.0059; p.Y895C, P = 0.0042; p.1067_1068del, P = 0.0006; p.N1215S, P = 0.0005; p.R1340C, P = 0.0003; p.E1723Q, P = 0.0012; p.R108C, P = 0.0041; p.E632K, P = 0.0131; p.K1051V, P = 0.0189; p.L1154V, P = 0.0004; p.E1163K, P = 0.0024; p.V1377E, P = 0.0291), c (p.G46D, P = 0.1542; p.R808C, P = 0.0274; p.Y895C, P = 0.0151; p.1067_1068del, P = 0.0062; p.N1215S, P = 0.1189; p.R1340C, P = 0.0244; p.E1723Q, P = 0.0385; p.R108C, P = 0.0187; p.E632K, P = 0.066; p.K1051V, P = 0.0536; p.L1154V, P = 0.0022; p.E1163K, P = 0.0304; p.V1377E, P = 0.1601) and d (p.G46D, P = 0.9917; p.R808C, P = 0.0103; p.Y895C, P = 0.005; p.1067_1068del, P = 0.0031; p.N1215S, P = 0.0944; p.R1340C, P = 0.0187; p.E1723Q, P = 0.4656; p.R108C, P = 0.0249; p.E632K, P = 0.0748; p.K1051V, P = 0.9471; p.L1154V, P = 0.0016; p.E1163K, P = 0.0194; p.V1377E, P = 0.9261). For panels b–d, two-tailed P-values were determined by the one-way ANOVA test followed by a Dunnett’s multiple-comparison test. *P < 0.05, **P < 0.01, and ***P < 0.001. EoE eosinophilic esophagitis, DSP desmoplakin, PPL periplakin, FITC fluorescein isothiocyanate, SEM standard error of the mean.

Fig. 4. Effect of DSP and PPL variants on RhoA activity.

a Active RhoA assays performed in EPC2 cells transduced with constructs encoding non-variant and mutated DSP or PPL. Statistics (versus non-variants): p.G46D, P = 0.132; p.R808C, P = 0.0691; p.Y895C, P = 0.0453; p.1067_1068del, P = 0.1013; p.N1215S, P = 0.2131; p.R1340C, P = 0.048; p.E1723Q, P = 0.0619; p.R108C, P = 0.4849; p.E632K, P = 0.0443; p.K1051V, P = 0.1589; p.L1154V, P = 0.0419; p.E1163K, P = 0.0035; p.V1377E, P = 0.8738. b Wound healing assays performed in EPC2 cells transduced with constructs encoding non-variant and mutated DSP (p.Y895C) treated with the Rho activator CN01 or the Rho kinase inhibitor Y27632. Quantification of the wound closure after 12 h was shown. Scale bar: 500 μM. Statistics: Non-variant vs. Variant, P = 0.0001; Non-variant vs. Variant + CN01, P = 0.7658; Non-variant vs. Variant + Y27632, P < 0.0001; Variant vs. Variant + CN01, P = 0.0002; Variant vs. Variant + Y27632, P = 0.0394; Variant + CN01 vs. Variant + Y27632, P < 0.0001. c Enzyme-linked immunosorbent assay of the level of active RhoA (RhoA-GTP) in protein lysates of biopsies from control individuals, non-familial patients with inactive EoE and non-familial patients with active EoE (Control, n = 3; Inactive EoE; n = 3; Active EoE, n = 5). Western blot analysis shows the expression of total RhoA in protein lysates of biopsies from each subject. Statistics: control vs. inactive EoE, P > 0.8803; control vs. active EoE, P = 0.0398; inactive EoE vs. active EoE, P = 0.0479. For panels a, b, data are representative of three experiments performed in duplicate and are presented as mean ± SEM. For panel c, data are presented as mean ± SEM, with markers representing biologically independent subjects. For panels a–c, two-tailed P-values were determined by the following tests: the one-way ANOVA test followed by a Dunnett’s multiple-comparison test (a) or Tukey’s multiple comparisons test (b, c). *P < 0.05, **P < 0.01, and ***P < 0.001. DSP desmoplakin, PPL periplakin, MW molecular weight, SEM standard error of the mean.

Fig. 5. Calpain 14–mediated desmosomal protein degradation.

a Immunoblots of lysates from DSP (non-variant or mutant p.Y895C)–transfected HEK293T cells with CAPN14 or enzymatically inactive CAPN14-C101A co-transfection for changes in DSP levels following the addition of exogenous Ca²⁺ (left). Protein remaining after activation of CAPN14 was determined by the difference in band intensity between after and before activation [(after x 100)/before] (right). Statistics: DSP WT vs. DSP mutant, P = 0.0131. b Effect of DSP and PPL variants on protein degradation on activation of co-transfected CAPN14. Total protein remaining was determined for each mutation. Total protein (%) = (each protein remaining x 100)/average of non-variant cells. Statistics (versus non-variants): p.G46D, P = 0.8212; p.R808C, P = 0.0296; p.Y895C, P = 0.0352; p.1067_1068del, P = 0.0125; p.N1215S, P = 0.2675; p.R1340C, P = 0.2581; p.E1723Q, P = 0.6186; p.R108C, P = 0.0712; p.E632K, P = 0.0156; p.K1051V, P = 0.0946; p.L1154V, P = 0.0323; p.E1163K, P = 0.2167; p.V1377E, P = 0.9157. c Gene expression of CAPN14 in human normal tissues from the Human Protein Atlas (https://www.proteinatlas.org/). d Calpain inhibition results in rescue of DSP and PPL levels by CAPN14-mediated degradation. Total protein (%) = (each protein remaining x 100)/average of cells without exogenous Ca²⁺ and SNJ-1945. Statistics: DSP (without SNJ-1945 vs. with SNJ-1945, P = 0.0028); PPL (without SNJ-1945 vs. with SNJ-1945, P = 0.0188). For panels a, b, and d, data are representative of three experiments performed in duplicate and are presented as mean ± SEM, and two-tailed P-values were determined by the unpaired t-test (a and d) or one-way ANOVA test followed by a Dunnett’s multiple-comparison test (b). *P < 0.05 and **P < 0.01. CAPN14 calpain-14, DSP desmoplakin, PPL periplakin, SEM standard error of the mean, WT wild-type.